Mangroves among the most carbon-rich forest in the tropics

Full text

LETTERS
PUBLISHED ONLINE: 3 APRIL 2011 | DOI: 10.1038/NGEO1123
Mangroves among the most carbon-rich forests in
the tropics
Daniel C. Donato1*, J. Boone Kauffman2, Daniel Murdiyarso3, Sofyan Kurnianto3, Melanie Stidham4
and Markku Kanninen5
Mangrove forests occur along ocean coastlines throughout the
tropics, and support numerous ecosystem services, including
fisheries production and nutrient cycling. However, the areal
extent of mangrove forests has declined by 30–50% over the
past half century as a result of coastal development, aqua-
culture expansion and over-harvesting1–4. Carbon emissions
resulting from mangrove loss are uncertain, owing in part to
a lack of broad-scale data on the amount of carbon stored
in these ecosystems, particularly below ground5. Here, we
quantified whole-ecosystem carbon storage by measuring tree
and dead wood biomass, soil carbon content, and soil depth in
25 mangrove forests across a broad area of the Indo-Pacific
region—spanning 30◦of latitude and 73◦of longitude—where
mangrove area and diversity are greatest4,6. These data indi-
cate that mangroves are among the most carbon-rich forests
in the tropics, containing on average 1,023 Mg carbon per
hectare. Organic-rich soils ranged from 0.5m to more than 3 m
in depth and accounted for 49–98% of carbon storage in these
systems. Combining our data with other published information,
we estimate that mangrove deforestation generates emissions
of 0.02–0.12 Pg carbon per year—as much as around 10% of
emissions from deforestation globally, despite accounting for
just 0.7% of tropical forest area6,7.
Deforestation and land-use change currently account for 8–20%
of global anthropogenic carbon dioxide (CO2) emissions, second
only to fossil fuel combustion7,8. Recent international climate
agreements highlight Reduced Emissions from Deforestation and
Degradation (REDD+) as a key and relatively cost-effective option
for mitigating climate change; the strategy aims to maintain
terrestrial carbon (C) stores through financial incentives for forest
conservation (for example, carbon credits). REDD+and similar
programs require rigorous monitoring of C pools and emissions8,9,
underscoring the importance of robust C storage estimates for
various forest types, particularly those with a combination of high
C density and widespread land-use change10.
Tropical wetland forests (for example, peatlands) contain
organic soils up to several metres deep and are among the largest
organic C reserves in the terrestrial biosphere11–13. Peatlands’
disproportionate importance in the link between land use and
climate change has received significant attention since 1997, when
peat fires associated with land clearing in Indonesia increased
atmospheric CO2enrichment by 13–40% over global annual
fossil fuel emissions11. This importance has prompted calls to
specifically address tropical peatlands in international climate
change mitigation strategies7,13.
1USDA Forest Service, Pacific Southwest Research Station, 60 Nowelo St., Hilo, Hawaii 96720, USA, 2USDA Forest Service, Northern Research Station, 271
Mast Rd., Durham, New Hampshire 03824, USA, 3Center for International Forestry Research (CIFOR), PO Box 0113 BOCBD, Bogor 16000, Indonesia,
4USDA Forest Service, International Programs, 1099 14th street NW, Suite 5500W, Washington, District of Columbia 20005, USA, 5Viikki Tropical
Resources Institute (VITRI), University of Helsinki, PO Box 27, FIN-00014, Finland. *e-mail:ddonato@wisc.edu.
Overlooked in this discussion are mangrove forests, which occur
along the coasts of most major oceans in 118 countries, adding
∼30–35% to the global area of tropical wetland forest over peat
swamps alone4,6,12. Renowned for an array of ecosystem services,
including fisheries and fibre production, sediment regulation, and
storm/tsunami protection2–4, mangroves are nevertheless declining
rapidly as a result of land clearing, aquaculture expansion,
overharvesting, and development2–6. A 30–50% areal decline over
the past half-century1,3 has prompted estimates that mangroves
may functionally disappear in as little as 100 years (refs 1,2). Rapid
twenty-first century sea-level rise has also been cited as a primary
threat to mangroves14, which have responded to past sea-level
changes by migrating landward or upward15.
Although mangroves are well known for high C assimilation
and flux rates16–22, data are surprisingly lacking on whole-ecosystem
carbon storage—the amount which stands to be released with
land-use conversion. Limited components of C storage have been
reported, most notably tree biomass17,18, but evidence of deep
organic-rich soils22–25 suggests these estimates miss the vast majority
of total ecosystem carbon. Mangrove soils consist of a variably
thick, tidally submerged suboxic layer (variously called ‘peat’ or
‘muck’) supporting anaerobic decomposition pathways and having
moderate to high C concentration16,20,21. Below-ground C storage
in mangrove soils is difficult to quantify5,21 and is not a simple
function of measured flux rates—it also integrates thousands of
years of variable deposition, transformation, and erosion dynamics
associated with fluctuating sea levels and episodic disturbances15.
No studies so far have integrated the necessary measurements for
total mangrove C storage across broad geographic domains.
In this study we quantified whole-ecosystem C storage in
mangroves across a broad tract of the Indo-Pacific region, the
geographic core of mangrove area (∼40% globally) and diversity4,6.
Study sites comprised wide variation in stand composition
and stature (Fig. 1, Supplementary Table S1), spanning 30◦
of latitude (8◦S–22◦N), 73◦of longitude (90◦–163◦E), and
including eastern Micronesia (Kosrae); western Micronesia
(Yap and Palau); Sulawesi, Java, Borneo (Indonesia); and the
Sundarbans (Ganges-Brahmaputra Delta, Bangladesh). Along
transects running inland from the seaward edge, we combined
established biometric techniques with soil coring to assess variations
in above- and below-ground C pools as a function of distance
from the seaward edge in two major geomorphic settings:
estuarine/river-delta and oceanic/fringe. Estuarine mangroves
(n=10) were situated on large alluvial deltas, often with a
protected lagoon; oceanic mangroves (n=15) were situated in
NATURE GEOSCIENCE |VOL 4 |MAY 2011 |www.nature.com/naturegeoscience 293
© 2011 Macmillan Publishers Limited. All rights reserved.

LETTERS NATURE GEOSCIENCE DOI: 10.1038/NGEO1123
a
b
Figure 1 |Examples of Indo-Pacific mangroves. The sample included a
broad range of stand stature, composition, and soil depth. a, Exemplary
large-stature, high-density mangrove dominated by Bruguiera, Borneo,
Indonesia (canopy height >15 m, canopy closure >90%, soil depth >3 m).
b, Exemplary small-stature, low-density mangrove dominated by
Rhizophora, Sulawesi, Indonesia (canopy height <4m, canopy closure
<60%, soil depth 0.35–0.78m). Both estuarine and oceanic mangroves
can exhibit both conditions (see Supplementary Table S1).
marine-edge settings, often the coasts of islands with fringing
coral reefs. Seaward distance and geomorphic setting may
influence C dynamics through differences in tidal flushing and
relative importance of allochthonous (river sediment) versus
autochthonous (in situ litter and root production) controls on soil
C accumulation5,16.
We found that mangroves are among the most C-dense forests
in the tropics (sample-wide mean: 1,023 Mg C ha−1±88 s.e.m.),
and exceptionally high compared to mean C storage of the
world’s major forest domains (Fig. 2). Estuarine sites contained
a mean of 1,074 Mg C ha−1(±171 s.e.m.); oceanic sites contained
990 ±96 Mg C ha−1. Above-ground C pools were sizeable (mean
159 Mg C ha−1, maximum 435 Mg C ha−1), but below-ground
storage in soils dominated, accounting for 71–98% and 49–90%
of total storage in estuarine and oceanic sites, respectively (Figs 2
and 3). Below-ground C storage was positively but weakly
correlated to above-ground storage (R2=0.21 and 0.50 in estuarine
and oceanic sites, respectively). Although soil C pools increased
slightly with distance from the seaward edge in oceanic sites
(because of increasing soil depth), changes in both above- and
Boreal Temperate Tropical
upland
Mangrove
Indo-Pacific
0
200
400
600
800
1,000
1,200
1,400
Ecosystem C storage (Mg ha¬1)
Above-ground live + dead
Soils 0¬30 cm depth + roots
Soils below 30 cm depth
Figure 2 |Comparison of mangrove C storage (mean ±95% confidence
interval) with that of major global forest domains. Mean C storage by
domain was derived from ref. 9, including default values for tree, litter, dead
wood, root:shoot ratios, and soils, with the assumption that the top 30cm
of soil contains 50% of all C residing in soil9, except for boreal forests
(25%). Domain means are presented for context; however some forest
types within each contain substantially higher or lower C stores9,10. In
general, the top 30 cm of soil C are considered the most vulnerable to
land-use change9; however in suboxic peat/muck soils, drainage,
excavation, and oxidation may influence deeper layers29.
below-ground C storage over this distance gradient were highly
variable and not statistically significant (Fig. 3).
So far, quantification of below-ground C storage in man-
groves has been impeded by a lack of concurrent data on soil
carbon concentration, bulk density, and depth, and how these
vary spatially5,21. We found high C concentration (% dry mass)
throughout the top metre of the soil profile, with a decrease
below 1 m (Fig. 4a). Carbon concentration was lower in es-
tuarine (mean =7.9%) versus oceanic (mean =14.6%) sites.
Soil bulk density (BD) did not differ significantly by setting or
distance from the seaward edge (generally ∼0.35–0.55 g cm−3),
but did increase with depth (Fig. 4b). Combining C concentra-
tion and BD yielded mean C densities of 0.038 g C cm−3and
0.061 g C cm−3in estuarine and oceanic soils, respectively. The
total depth of the peat/muck layer differed between estuarine
and oceanic sites (Fig. 4c) and was the main driver of varia-
tions in below-ground C storage (Fig. 3). Estuarine stands over-
lie deep alluvial sediment deposits, usually exceeding 3 m depth;
oceanic stands contained a distinct organic-rich layer overlying
hard coral sand or rock, with peat/muck thickness increasing
from a mean of 1.2 m (±0.2 s.e.m.) near the seaward edge to
1.7 m (±0.2 s.e.m.) 135 m inland (Fig. 4c). In terms of total
below-ground C storage, the shallower soil depth in oceanic man-
groves was compensated in part by higher soil organic C con-
centration (Fig. 4a,c).
These data indicate that high productivity and C flux rates
in mangroves16–22 are indeed accompanied by high C storage,
especially below ground. High per-hectare C storage coupled with
a pan-tropical distribution (total area ∼14 million ha; refs 4,6)
suggests mangroves are a globally important surface C reserve.
Although our sample is not intended to represent all mangrove
types (precluding simple scaling up), some constraints on global
storage can be derived by combining an uncertainty range from
our empirical data (5th to 95th percentile C storage values) with
additional global data on soil C concentration, depth, and standing
biomass16,17,21,23,24 (see Methods in Supplementary Information).
This approach yields an estimate of 4.0–20 Pg C globally. This
estimate will undoubtedly be refined, but suggests mangroves add
significantly to tropical wetland forest C storage (for example,
tropical peatlands: ∼82–92 Pg C; ref. 12).
294 NATURE GEOSCIENCE |VOL 4 |MAY 2011 |www.nature.com/naturegeoscience
© 2011 Macmillan Publishers Limited. All rights reserved.

NATURE GEOSCIENCE DOI: 10.1038/NGEO1123 LETTERS
400
200
0
200
400
600
800
1,000
1,200
1,400
Below-ground Above-ground
C storage (Mg ha¬1)
400
200
0
200
400
600
800
1,000
1,200
1,400
Below-ground Above-ground
C storage (Mg ha¬1)
0 10 35 60 85 110 135
Distance from seaward edge (m)
0 10 35 60 85 110 135
Distance from seaward edge (m)
Estuarine mangroves Oceanic mangroves
Trees
Down wood
Roots
Soil
ab
Figure 3 |Above- and below-ground C pools in Indo-Pacific mangroves, assessed by distance from the seaward edge. a, Estuarine mangroves situated on
large alluvial deltas. b, Oceanic mangroves situated in marine edge environments—for example, island coasts. Below-ground C comprised 71–98% and
49–90% of ecosystem C in estuarine and oceanic sites, respectively. Overall carbon storage did not vary significantly with distance from the seaward edge
in either setting over the range sampled (P>0.10 for above-ground, below-ground, and total C storage by functional data analysis (FDA, see Methods);
95% CIs for rates-of-change all overlapped zero and were between −1.2 and 3.9 Mg C ha−1per metre of distance from edge).
Carbon emissions from land-use change in mangroves are
not well understood. Our data suggest a potential for large
emissions owing to perturbation of large C stocks. The fate of
below-ground pools is particularly understudied, but available
evidence suggests that clearing, drainage, and/or conversion
to aquaculture—aside from affecting vegetation biomass—also
decreases mangrove soil C significantly16,22,26–28. In upland forests,
the top 30 cm of soil are generally considered the most susceptible
to land-use change9; however in wetland forests, drainage and
oxidation of formerly suboxic soils may also influence deeper
layers29. To provide some constraints on estimated emissions,
we used a similar uncertainty propagation technique, combining
our C storage values with other global data16,17 and applying
a range of assumptions regarding land-use effects on above-
and below-ground pools (see Supplementary Information). This
approach yields a plausible estimate of 112–392 Mg C released
per hectare cleared, depending in large part on how deeply soil
C is affected by different land uses. Coupled with published
ranges of mangrove deforestation rate (1–2%; refs 1,4) and global
area (13.7–15.2 million ha; refs 4,6), this estimate leads to global
emissions on the order of 0.02–0.12 Pg C yr−1. This rate adds
significantly to oft-cited peatland emissions (0.30 Pg C yr−1) and
global deforestation emissions (∼1.2 Pg C yr−1; ref. 7) despite
accounting only for loss of standing stocks but not other known
mangrove-conversion influences, such as decreased C sequestration
rate, burial efficiency, and export to ocean16,18, nor increases in
normally-low methanogenesis in some disturbed soils16,27.
In addition to direct losses of forest cover, land-use activities
will also impact mangrove responses to sea-level rise14,15. Man-
groves have been remarkably persistent through rapid sea-level rises
(5–15 mm yr−1) during the late Quaternary Period (0–18,000 yr bp)
because of (1) landward migration, and (2) autogenic changes
in soil-surface elevation through below-ground organic matter
production and/or sedimentation15. Under current climate trends,
sea level is projected to rise 18–79 cm from 1999–2099 (higher
if ice-sheet melting continues accelerating)8,30, implying a period-
averaged rate of ∼1.8–7.9 mm yr−1, notwithstanding local varia-
tions and temporal nonlinearities. Although this rate is not unprece-
dented, it is unclear yet whether mangroves are currently keeping
pace with sea levels14,15. Anthropogenic influences could constrain
future resilience to sea-level rise through coastal developments
that impede inland migration (for example, roads, infrastructure),
upland land uses that alter sediment and water inputs (for example,
dams, land clearing), and mangrove degradation that reduces
below-ground productivity14. This synergy of land use and climate
change impacts presents additional uncertainties for the fate and
management of coastal C stores.
Critical uncertainties remain before estimates of mangrove C
storage and land-use emissions can be improved. Among these are
geographic variations in soil depth, a key but unknown parameter
in most regions5,21. Similarly, empirical data on land-use change
impacts on soil C is strongly lacking, especially for deep layers
(but see refs 26–28). Quantitative estimates are also needed of
the relative area occupied by estuarine/delta and oceanic/fringe
mangroves, which is not addressed in most analyses of mangrove
area4,6. Because these two systems store below-ground C differently,
improved spatial data will greatly refine estimates of global C storage
and emissions owing to disturbance.
Our data show that discussion of the key role of tropical wetland
forests in climate change could be broadened significantly to include
mangroves. Southeast Asian peatlands are currently being advanced
as an essential component of climate change mitigation strategies
such as REDD+(refs 7,13), and mangroves share many of the
same relevant characteristics: deep organic-rich soils, exceptionally
high C storage, and extensive deforestation/degradation resulting in
potentially large greenhouse gas emissions. The well-known ecosys-
tem services and geographic distribution of mangroves1–4 suggest
these mitigation strategies could be effective in providing ancillary
benefits as well as potential REDD+ opportunities in many tropical
countries. Because land use in mangroves affects not only standing
stocks but also ecosystem response to sea-level rise, maintaining
these C stores will require both in situ mitigation (for example,
reducing conversion rates) as well as facilitating adaptation to
rising seas. The latter challenge is largely unique to management
NATURE GEOSCIENCE |VOL 4 |MAY 2011 |www.nature.com/naturegeoscience 295
© 2011 Macmillan Publishers Limited. All rights reserved.

LETTERS NATURE GEOSCIENCE DOI: 10.1038/NGEO1123
0 10 35 60 85 110 135
Distance from seaward edge (m)
0 10 35 60 85 110 135
Distance from seaward edge (m)
Organic C content (%)
0
0
1
2
10 20
Depth (m)
0 10 20 0 10 20 0 10 20 0 10 20 0 10 20
Estuarine Oceanic
Bulk density (g cm¬3)
0.3 0.6 0.3 0.6 0.3 0.6 0.3 0.6 0.3 0.6 0.3 0.6
Estuarine/oceanic combined
Soil depth (m)
Estuarine (88% >3 m)
Oceanic
0
1
2
Depth (m)
Soil depth (m)
3
2
1
0
a
b
c
Figure 4 |Soil properties determining below-ground carbon storage in
Indo-Pacific mangroves. a, Soil C concentration was greater in oceanic
(mean =14.6%) versus estuarine (mean =7.9%) sites (P=0.001), and
decreased with depth (P<0.0001; effect stronger in oceanic sites).
Change in C concentration with seaward distance was biologically
insignificant. b, Soil bulk density did not differ significantly with setting
(P=0.79); hence one line is shown combining both settings. Bulk density
increased with depth (P<0.0001) but not seaward distance (P=0.20),
and a distance*depth interaction term was not significant (P=0.47). c, Soil
depth increased with distance from the seaward edge in oceanic stands
(FDA result: P=0.002, 95% CI for rate-of-change =21–65 cm increase
per 100 m distance).
of coastal forests, calling for watershed-scale approaches, such
as landscape buffers for accommodating inland migration where
possible, maintenance of critical upstream sediment inputs, and
addressing degradation of mangrove productivity from pollution
and other exogenous impacts14,15.
Methods
We sampled 25 mangrove sites (n=10 estuarine, n=15 oceanic) across the
Indo-Pacific (8◦S–22◦N, 90◦–163◦E) using a transect starting from, and running
perpendicular to, the seaward edge. To maximize scope and representativeness,
we stratified the sample across a broad range of stand conditions—including
small-stature stands and shallow soils (<4 m canopy height, <10 cm mean tree
diameter, <0.5 m soil depth) to large-stature stands and deep soils (>15 m canopy
height, >20 cm mean tree diameter, soil depth >3m) (Supplementary Table S1).
These structural characteristics of forest stature and soil depth are primary
determinants of C storage, probably more so than environmental gradients or
geographic variation per se. Specific transect starts were determined randomly
a priori from aerial imagery, notwithstanding constraints of access and land
ownership. Within six circular sample plots spaced at 25-m intervals along each
transect, we measured standing tree and down wood (dead wood on forest floor)
biomass using standard biometric techniques (stem surveys, planar intercept
transects), then applying region-specific or common allometric equations and
C:biomass conversions for both above-ground and below-ground biomass. We
measured soil depths at three systematic locations in each plot using a graduated
aluminium probe (inference limit 3 m). We extracted a soil core from each plot
using a 6.4-cm open-face peat auger to minimize sample disturbance/compaction,
systematically divided the soil profile into depth intervals, and collected subsamples
from each interval. Subsamples were dried to constant mass and weighed for
bulk density determination, then analysed for C concentration using the dry
combustion method (induction furnace). Standard error in total ecosystem C
storage was computed by propagating standard errors of component pools. For
estuarine and oceanic sites separately, we analysed changes in soil depth and C
pools with distance from the seaward edge using functional data analysis (site-level
regressions for rate-of-change with distance, followed by a one-sample parametric
test on all rates-of-change for strength of positive or negative relationship). We
analysed spatial variations in soil C concentration and bulk density using linear
mixed-effects regression models, assessing fixed effects of depth, distance from
the seaward edge, and geomorphic setting, with a random effect of site to account
for within-site dependence. Ranges for global C storage and emission rates were
obtained using 5th percentile, mean, or 95th percentile estimates from this study
(which accounts for the possibility that biomass and soil C pools differ globally
from our mean values—higher or lower), with an adjusted soil C density based
on a recent global analysis16, combined with recently published low to high
estimates of global mangrove area and deforestation rate1,4,6. See full Methods in
Supplementary Information.
Received 30 September 2010; accepted 23 February 2011;
published online 3 April 2011
References
1. Duke, N. C. et al. A world without mangroves? Science 317, 41–42 (2007).
2. Polidoro, B. A. et al. The loss of species: Mangrove extinction risk and
geographic areas of global concern. PLoS ONE 5, e10095 (2010).
3. Alongi, D. M. Present state and future of the world’s mangrove forests.
Environ. Conserv. 29, 331–349 (2002).
4. Food and Agriculture Organization of the United Nations (FAO). The World’s
Mangroves 1980–2005 (FAO Forestry Paper 153. FAO, 2007).
5. Bouillon, S., Rivera-Monroy, V. H., Twilley, R. R., Kairo, J. G. & Mangroves,
in The Management of Natural Coastal Carbon Sinks (eds Laffoley, D. &
Grimsditch) (IUCN, 2009).
6. Giri, C. et al. Status and distribution of mangrove forests of the world using
earth observation satellite data. Glob. Ecol. Biogeogr. 20, 154–159 (2011).
7. van der Werf, G. R. et al. CO2emissions from forest loss. Nature Geosci. 2,
737–738 (2009).
8. IPCC in The Fourth Assessment Report Climate Change 2007 (eds Pachauri, R.
K. & Reisinger, A.) (IPCC, 2007).
9. Intergovernmental Panel on Climate Change (IPCC) in Good Practice Guidance
for Land use, Land-use Change, and Forestry (eds Penman, J. et al.) (Institute
for Global Environmental Strategies, 2003).
10. Keith, H., Mackey, B. G. & Lindenmayer, D. B. Re-evaluation of forest biomass
carbon stocks and lessons from the world’s most carbon-dense forests.
Proc. Natl Acad. Sci. USA 106, 11635–11640 (2009).
11. Page, S. E. et al. The amount of carbon released from peat and forest fires in
Indonesia during 1997. Nature 420, 61–65 (2002).
12. Page, S. E., Rieley, J. O. & Banks, C. J. Global and regional importance of the
tropical peatland carbon pool. Glob. Change Biol. 17, 798–818 (2011).
13. Murdiyarso, D. M., Hergoualc’h, K. & Verchot, L. V. Opportunities for
reducing greenhouse gas emissions in tropical peatlands. Proc. Natl Acad.
Sci. USA 107, 19655–19660 (2010).
14. Gilman, E. L., Ellison, J., Duke, N. C. & Field, C. Threats to mangroves from
climate change and adaptation options. Aquat. Bot. 89, 237–250 (2008).
15. Alongi, D. M. Mangrove forests: Resilience, protection from tsunamis, and
responses to global climate change. Estuar. Coast Shelf Sci. 76, 1–13 (2008).
16. Kristensen, E., Bouillon, S., Dittmar, T. & Marchand, C. Organic carbon
dynamics in mangrove ecosystems. Aquat. Bot. 89, 201–219 (2008).
17. Komiyama, A., Ong, J. E. & Poungparn, S. Allometry, biomass, and productivity
of mangrove forests. Aquat. Bot. 89, 128–137 (2008).
18. Twilley, R. R., Chen, R. H. & Hargis, T. Carbon sinks in mangroves and
their implications to carbon budget of tropical coastal ecosystems. Water Air
Soil Pollut. 64, 265–288 (1992).
19. Bouillon, S. et al. Mangrove production and carbon sinks: A revision of global
budget estimates. Glob. Biogeochem. Cycles 22, GB2013 (2008).
20. Alongi, D. M. et al. Sediment accumulation and organic material flux in a
managed mangrove ecosystem: Estimates of land–ocean–atmosphere exchange
in peninsular Malaysia. Mar. Geol. 208, 383–402 (2004).
21. Chmura, G. L., Anisfeld, S. C., Cahoon, D. R. & Lynch, J. C. Global carbon
sequestration in tidal, saline wetland soils. Glob. Biogeochem. Cycles 17,
1111 (2003).
22. Eong, O. J. Mangroves—a carbon source and sink. Chemosphere 27,
1097–1107 (1993).
23. Golley, F., Odum, H. T. & Wilson, R. F. The structure and metabolism of a
Puerto Rican red mangrove forest in May. Ecology 43, 9–19 (1962).
24. Fujimoto, K. et al. Belowground carbon storage of Micronesian mangrove
forests. Ecol. Res. 14, 409–413 (1999).
25. Matsui, N. Estimated stocks of organic carbon in mangrove roots and
sediments in Hinchinbrook Channel, Australia. Mangr. Salt Marsh. 2,
199–204 (1998).
26. Sjöling, S., Mohammed, S. M., Lyimo, T. J. & Kyaruzi, J. J. Benthic bacterial
diversity and nutrient processes in mangroves: Impact of deforestation.
Estuar. Coast Shelf Sci. 63, 397–406 (2005).
296 NATURE GEOSCIENCE |VOL 4 |MAY 2011 |www.nature.com/naturegeoscience
© 2011 Macmillan Publishers Limited. All rights reserved.

NATURE GEOSCIENCE DOI: 10.1038/NGEO1123 LETTERS
27. Strangmann, A., Bashan, Y. & Giani, L. Methane in pristine and impaired
mangrove soils and its possible effects on establishment of mangrove seedlings.
Biol. Fertil. Soils 44, 511–519 (2008).
28. Granek, E. & Ruttenberg, B. I. Changes in biotic and abiotic processes following
mangrove clearing. Estuar. Coast Shelf Sci. 80, 555–562 (2008).
29. Hooijer, A., Silvius, M., Wösten, H. & Page, S. PEAT-CO2:Assessment of
CO2Emissions from Drained Peatlands in SE Asia. Delft Hydraulics Report
Q3943 (2006).
30. Church, J. A. et al. Understanding global sea levels: Past, present and future.
Sustain. Sci. 3, 9–22 (2008).
Acknowledgements
We thank our many international partners and field personnel for assistance with
logistics and data collection: Kosrae Island Resource Management Authority; Yap
State Forestry; Orangutan Foundation International; Indonesian Directorate General
for Forest Protection and Nature Conservation; University of Manado and Bogor
Agricultural University, Indonesia; Bangladesh Forest Department; and KPSKSA
(Cilacap, Indonesia). We thank K. Gerow for statistical assistance, and R. Mackenzie,
C. Kryss and J. Bonham for assistance compiling site data. Funding was provided by
USDA Forest Service International Programs and the Australian Agency for International
Development (AusAID).
Author contributions
D.C.D. co-designed the study, collected field data, performed data analyses, and led the
writing of the paper. J.B.K. conceived and co-designed the study, and contributed to data
collection and writing. D.M. co-conceived the study, arranged for and contributed to data
collection, and contributed to writing. S.K. contributed to data collection, data analysis,
and writing. M.S. collected field data and contributed to writing. M.K. co-conceived the
study and contributed to writing.
Additional information
The authors declare no competing financial interests. Supplementary information
accompanies this paper on www.nature.com/naturegeoscience. Reprints and permissions
information is available online at http://npg.nature.com/reprintsandpermissions.
Correspondence and requests for materials should be addressed to D.C.D.
NATURE GEOSCIENCE |VOL 4 |MAY 2011 |www.nature.com/naturegeoscience 297
© 2011 Macmillan Publishers Limited. All rights reserved.